人脸生成（Face Generation）¶

在该项目中，你将使用生成式对抗网络（Generative Adversarial Nets）来生成新的人脸图像。

获取数据¶

该项目将使用以下数据集：

MNIST
CelebA

由于 CelebA 数据集比较复杂，而且这是你第一次使用 GANs。我们想让你先在 MNIST 数据集上测试你的 GANs 模型，以让你更快的评估所建立模型的性能。

如果你在使用 FloydHub, 请将 data_dir 设置为 "/input" 并使用 FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

探索数据（Explore the Data）¶

MNIST¶

MNIST 是一个手写数字的图像数据集。你可以更改 show_n_images 探索此数据集。

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7f4bac18b358>

CelebA¶

CelebFaces Attributes Dataset (CelebA) 是一个包含 20 多万张名人图片及相关图片说明的数据集。你将用此数据集生成人脸，不会用不到相关说明。你可以更改 show_n_images 探索此数据集。

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7f4bac069438>

预处理数据（Preprocess the Data）¶

由于该项目的重点是建立 GANs 模型，我们将为你预处理数据。

经过数据预处理，MNIST 和 CelebA 数据集的值在 28×28 维度图像的 [-0.5, 0.5] 范围内。CelebA 数据集中的图像裁剪了非脸部的图像部分，然后调整到 28x28 维度。

MNIST 数据集中的图像是单通道的黑白图像，CelebA 数据集中的图像是三通道的 RGB 彩色图像。

建立神经网络（Build the Neural Network）¶

你将通过部署以下函数来建立 GANs 的主要组成部分:

model_inputs
discriminator
generator
model_loss
model_opt
train

检查 TensorFlow 版本并获取 GPU 型号¶

检查你是否使用正确的 TensorFlow 版本，并获取 GPU 型号

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.0.1
Default GPU Device: /gpu:0

输入（Input）¶

部署 model_inputs 函数以创建用于神经网络的占位符 (TF Placeholders)。请创建以下占位符：

输入图像占位符: 使用 image_width，image_height 和 image_channels 设置为 rank 4。
输入 Z 占位符: 设置为 rank 2，并命名为 z_dim。
学习速率占位符: 设置为 rank 0。

返回占位符元组的形状为 (tensor of real input images, tensor of z data, learning rate)。

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    inputs_real = tf.placeholder(tf.float32, (None, image_height, image_width,image_channels), name='input_real') 
    inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')
    learning_rate = tf.placeholder(tf.float32,name="learning_rate")
    return inputs_real, inputs_z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

辨别器（Discriminator）¶

部署 discriminator 函数创建辨别器神经网络以辨别 images。该函数应能够重复使用神经网络中的各种变量。在 tf.variable_scope 中使用 "discriminator" 的变量空间名来重复使用该函数中的变量。

该函数应返回形如 (tensor output of the discriminator, tensor logits of the discriminator) 的元组。

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    alpha = 0.2
    with tf.variable_scope('discriminator',reuse=reuse):
        # Input layer is 32x32x3
        x1 = tf.layers.conv2d(images, 64, 5, strides=2, padding='same')
        relu1 = tf.maximum(alpha * x1, x1)
        # 16x16x64
        
        x2 = tf.layers.conv2d(relu1, 128, 5, strides=2, padding='same')
        bn2 = tf.layers.batch_normalization(x2, training=True)
        relu2 = tf.maximum(alpha * bn2, bn2)
        # 8x8x128
        
        x3 = tf.layers.conv2d(relu2, 256, 5, strides=2, padding='same')
        bn3 = tf.layers.batch_normalization(x3, training=True)
        relu3 = tf.maximum(alpha * bn3, bn3)
        # 4x4x256

        # Flatten it
        flat = tf.reshape(relu3, (-1, 4*4*256))
        logits = tf.layers.dense(flat, 1)
        out = tf.sigmoid(logits)
        
        return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

生成器（Generator）¶

部署 generator 函数以使用 z 生成图像。该函数应能够重复使用神经网络中的各种变量。在 tf.variable_scope 中使用 "generator" 的变量空间名来重复使用该函数中的变量。

该函数应返回所生成的 28 x 28 x out_channel_dim 维度图像。

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    reuse = not is_train
    alpha = 0.1
    with tf.variable_scope('generator', reuse=reuse):
        # First fully connected layer
        x1 = tf.layers.dense(z, 4*4*512)
        # Reshape it to start the convolutional stack
        x1 = tf.reshape(x1, (-1, 4, 4, 512))
        x1 = tf.layers.batch_normalization(x1, training=is_train)
        x1 = tf.maximum(alpha * x1, x1)
        # 4x4x512 now
        
        x2 = tf.layers.conv2d_transpose(x1, 256, 4, strides=1, padding='valid')
        x2 = tf.layers.batch_normalization(x2, training=is_train)
        x2 = tf.maximum(alpha * x2, x2)
        # 7x7x256 now

        x3 = tf.layers.conv2d_transpose(x2, 128, 5, strides=2, padding='same')
        x3 = tf.layers.batch_normalization(x3, training=is_train)
        x3 = tf.maximum(alpha * x3, x3)
        # 14x14x128

        # Output layer
        logits = tf.layers.conv2d_transpose(x3, out_channel_dim, 5, strides=2, padding='same')
        out = tf.tanh(logits)
        
        return out

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

损失函数（Loss）¶

部署 model_loss 函数训练并计算 GANs 的损失。该函数应返回形如 (discriminator loss, generator loss) 的元组。

使用你已实现的函数：

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """

    g_model = generator(input_z, out_channel_dim)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)

    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake

    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

优化（Optimization）¶

部署 model_opt 函数实现对 GANs 的优化。使用 tf.trainable_variables 获取可训练的所有变量。通过变量空间名 discriminator 和 generator 来过滤变量。该函数应返回形如 (discriminator training operation, generator training operation) 的元组。

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # Get weights and bias to update
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]

    # Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)

    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

训练神经网络（Neural Network Training）¶

输出显示¶

使用该函数可以显示生成器 (Generator) 在训练过程中的当前输出，这会帮你评估 GANs 模型的训练程度。

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

训练¶

部署 train 函数以建立并训练 GANs 模型。记得使用以下你已完成的函数：

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

使用 show_generator_output 函数显示 generator 在训练过程中的输出。

注意：在每个批次 (batch) 中运行 show_generator_output 函数会显著增加训练时间与该 notebook 的体积。推荐每 100 批次输出一次 generator 的输出。

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    #saver = tf.train.Saver()
    losses = []
    _,image_width,image_height,image_channels = data_shape
    input_real, input_z , _ = model_inputs(image_width, image_height, image_channels, z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, image_channels)
    d_opt, g_opt = model_opt(d_loss, g_loss,learning_rate, beta1)
    steps = 0
    n_images = 20
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                # TODO: Train Model
                steps += 1
                batch_images = batch_images * 2

                # Sample random noise for G
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))

                # Run optimizers
                _ = sess.run(d_opt, feed_dict={input_real: batch_images, input_z: batch_z})
                _ = sess.run(g_opt, feed_dict={input_z: batch_z, input_real: batch_images})

                if steps % 10 == 0:
                    # At the end of each epoch, get the losses and print them out
                    train_loss_d = d_loss.eval({input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_real: batch_images, input_z: batch_z})

                    print("Epoch {}/{}...".format(epoch_i+1, epoch_count),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))
                    # Save losses to view after training
                    losses.append((train_loss_d, train_loss_g))

                if steps % 100 == 0:
                    show_generator_output(sess, n_images, input_z, image_channels, data_image_mode)

MNIST¶

在 MNIST 上测试你的 GANs 模型。经过 2 次迭代，GANs 应该能够生成类似手写数字的图像。确保生成器 (generator) 低于辨别器 (discriminator) 的损失，或接近 0。

batch_size = 32
z_dim = 100
learning_rate = 0.0002
beta1 = 0.25


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 1/2... Discriminator Loss: 4.1405... Generator Loss: 0.0236
Epoch 1/2... Discriminator Loss: 3.2912... Generator Loss: 0.0468
Epoch 1/2... Discriminator Loss: 3.2491... Generator Loss: 0.0576
Epoch 1/2... Discriminator Loss: 0.5010... Generator Loss: 1.8376
Epoch 1/2... Discriminator Loss: 0.2218... Generator Loss: 2.8362
Epoch 1/2... Discriminator Loss: 1.6584... Generator Loss: 7.8315
Epoch 1/2... Discriminator Loss: 1.4870... Generator Loss: 0.3574
Epoch 1/2... Discriminator Loss: 1.5501... Generator Loss: 0.8129
Epoch 1/2... Discriminator Loss: 0.7523... Generator Loss: 2.0181
Epoch 1/2... Discriminator Loss: 1.0027... Generator Loss: 0.5933

Epoch 1/2... Discriminator Loss: 1.8306... Generator Loss: 0.2827
Epoch 1/2... Discriminator Loss: 1.1922... Generator Loss: 0.6582
Epoch 1/2... Discriminator Loss: 1.9662... Generator Loss: 0.2172
Epoch 1/2... Discriminator Loss: 1.6523... Generator Loss: 0.3030
Epoch 1/2... Discriminator Loss: 2.0579... Generator Loss: 0.1766
Epoch 1/2... Discriminator Loss: 1.4252... Generator Loss: 0.3999
Epoch 1/2... Discriminator Loss: 1.4720... Generator Loss: 0.3288
Epoch 1/2... Discriminator Loss: 1.4703... Generator Loss: 0.3536
Epoch 1/2... Discriminator Loss: 1.4313... Generator Loss: 0.3722
Epoch 1/2... Discriminator Loss: 1.2222... Generator Loss: 0.5137

Epoch 1/2... Discriminator Loss: 1.1885... Generator Loss: 0.5226
Epoch 1/2... Discriminator Loss: 1.1504... Generator Loss: 0.5755
Epoch 1/2... Discriminator Loss: 1.0867... Generator Loss: 1.2582
Epoch 1/2... Discriminator Loss: 1.2706... Generator Loss: 1.5405
Epoch 1/2... Discriminator Loss: 1.1147... Generator Loss: 0.9159
Epoch 1/2... Discriminator Loss: 1.1817... Generator Loss: 1.3892
Epoch 1/2... Discriminator Loss: 0.9704... Generator Loss: 0.9999
Epoch 1/2... Discriminator Loss: 0.9588... Generator Loss: 0.8352
Epoch 1/2... Discriminator Loss: 1.1735... Generator Loss: 0.4929
Epoch 1/2... Discriminator Loss: 0.9723... Generator Loss: 0.8836

Epoch 1/2... Discriminator Loss: 1.4125... Generator Loss: 0.3943
Epoch 1/2... Discriminator Loss: 0.8212... Generator Loss: 1.1417
Epoch 1/2... Discriminator Loss: 0.8619... Generator Loss: 1.1297
Epoch 1/2... Discriminator Loss: 1.1462... Generator Loss: 0.4955
Epoch 1/2... Discriminator Loss: 1.0523... Generator Loss: 0.6219
Epoch 1/2... Discriminator Loss: 1.0620... Generator Loss: 0.5816
Epoch 1/2... Discriminator Loss: 0.9124... Generator Loss: 0.9265
Epoch 1/2... Discriminator Loss: 0.9116... Generator Loss: 1.0457
Epoch 1/2... Discriminator Loss: 0.9971... Generator Loss: 1.6784
Epoch 1/2... Discriminator Loss: 1.3522... Generator Loss: 0.3914

Epoch 1/2... Discriminator Loss: 1.2697... Generator Loss: 1.4227
Epoch 1/2... Discriminator Loss: 0.9827... Generator Loss: 0.9020
Epoch 1/2... Discriminator Loss: 1.1887... Generator Loss: 0.5222
Epoch 1/2... Discriminator Loss: 1.2819... Generator Loss: 0.4090
Epoch 1/2... Discriminator Loss: 1.2498... Generator Loss: 1.6521
Epoch 1/2... Discriminator Loss: 0.9622... Generator Loss: 1.2218
Epoch 1/2... Discriminator Loss: 0.9304... Generator Loss: 0.8471
Epoch 1/2... Discriminator Loss: 1.3151... Generator Loss: 0.4071
Epoch 1/2... Discriminator Loss: 1.1079... Generator Loss: 1.2835
Epoch 1/2... Discriminator Loss: 1.0043... Generator Loss: 1.6226

Epoch 1/2... Discriminator Loss: 1.2249... Generator Loss: 0.4734
Epoch 1/2... Discriminator Loss: 1.0646... Generator Loss: 1.9693
Epoch 1/2... Discriminator Loss: 1.0924... Generator Loss: 1.4315
Epoch 1/2... Discriminator Loss: 0.9532... Generator Loss: 0.7600
Epoch 1/2... Discriminator Loss: 1.2131... Generator Loss: 0.4902
Epoch 1/2... Discriminator Loss: 1.1842... Generator Loss: 1.8068
Epoch 1/2... Discriminator Loss: 1.3072... Generator Loss: 0.4830
Epoch 1/2... Discriminator Loss: 0.9788... Generator Loss: 0.8153
Epoch 1/2... Discriminator Loss: 1.4211... Generator Loss: 0.3730
Epoch 1/2... Discriminator Loss: 1.4874... Generator Loss: 0.3688

Epoch 1/2... Discriminator Loss: 0.9356... Generator Loss: 1.3109
Epoch 1/2... Discriminator Loss: 1.6871... Generator Loss: 0.2658
Epoch 1/2... Discriminator Loss: 1.4989... Generator Loss: 0.3349
Epoch 1/2... Discriminator Loss: 1.1462... Generator Loss: 0.5941
Epoch 1/2... Discriminator Loss: 0.8951... Generator Loss: 1.1002
Epoch 1/2... Discriminator Loss: 1.0558... Generator Loss: 0.7076
Epoch 1/2... Discriminator Loss: 1.6089... Generator Loss: 0.2866
Epoch 1/2... Discriminator Loss: 1.2123... Generator Loss: 0.9849
Epoch 1/2... Discriminator Loss: 1.0511... Generator Loss: 1.2001
Epoch 1/2... Discriminator Loss: 1.0266... Generator Loss: 0.6969

Epoch 1/2... Discriminator Loss: 0.9219... Generator Loss: 1.6011
Epoch 1/2... Discriminator Loss: 0.9269... Generator Loss: 1.3097
Epoch 1/2... Discriminator Loss: 1.1491... Generator Loss: 0.5199
Epoch 1/2... Discriminator Loss: 1.2876... Generator Loss: 1.5543
Epoch 1/2... Discriminator Loss: 0.9575... Generator Loss: 1.3531
Epoch 1/2... Discriminator Loss: 1.4547... Generator Loss: 0.3447
Epoch 1/2... Discriminator Loss: 0.9855... Generator Loss: 0.7543
Epoch 1/2... Discriminator Loss: 1.3626... Generator Loss: 0.4407
Epoch 1/2... Discriminator Loss: 1.3603... Generator Loss: 0.3770
Epoch 1/2... Discriminator Loss: 0.9237... Generator Loss: 1.7581

Epoch 1/2... Discriminator Loss: 0.9180... Generator Loss: 0.8215
Epoch 1/2... Discriminator Loss: 1.2140... Generator Loss: 0.4610
Epoch 1/2... Discriminator Loss: 1.0226... Generator Loss: 0.6382
Epoch 1/2... Discriminator Loss: 1.1976... Generator Loss: 0.5333
Epoch 1/2... Discriminator Loss: 1.0848... Generator Loss: 1.0949
Epoch 1/2... Discriminator Loss: 1.1080... Generator Loss: 0.5784
Epoch 1/2... Discriminator Loss: 2.6112... Generator Loss: 0.1001
Epoch 1/2... Discriminator Loss: 0.9289... Generator Loss: 0.9370
Epoch 1/2... Discriminator Loss: 1.0453... Generator Loss: 0.6451
Epoch 1/2... Discriminator Loss: 1.2218... Generator Loss: 0.4482

Epoch 1/2... Discriminator Loss: 0.9493... Generator Loss: 0.7937
Epoch 1/2... Discriminator Loss: 0.9021... Generator Loss: 1.4691
Epoch 1/2... Discriminator Loss: 1.7453... Generator Loss: 2.4360
Epoch 1/2... Discriminator Loss: 0.8609... Generator Loss: 0.9861
Epoch 1/2... Discriminator Loss: 2.2252... Generator Loss: 0.1485
Epoch 1/2... Discriminator Loss: 1.1839... Generator Loss: 0.5230
Epoch 1/2... Discriminator Loss: 0.8383... Generator Loss: 1.5251
Epoch 1/2... Discriminator Loss: 1.0183... Generator Loss: 0.6223
Epoch 1/2... Discriminator Loss: 1.4023... Generator Loss: 0.3662
Epoch 1/2... Discriminator Loss: 1.9718... Generator Loss: 0.1789

CelebA¶

在 CelebA 上运行你的 GANs 模型。在一般的GPU上运行每次迭代大约需要 20 分钟。你可以运行整个迭代，或者当 GANs 开始产生真实人脸图像时停止它。

batch_size = 16
z_dim = 100
learning_rate = 0.0003
beta1 = 0.3


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 1/1... Discriminator Loss: 4.5888... Generator Loss: 0.0135
Epoch 1/1... Discriminator Loss: 4.0115... Generator Loss: 0.0293
Epoch 1/1... Discriminator Loss: 1.8160... Generator Loss: 9.3494
Epoch 1/1... Discriminator Loss: 4.9376... Generator Loss: 0.0110
Epoch 1/1... Discriminator Loss: 0.6582... Generator Loss: 2.0976
Epoch 1/1... Discriminator Loss: 2.6202... Generator Loss: 0.0983
Epoch 1/1... Discriminator Loss: 0.6539... Generator Loss: 6.4723
Epoch 1/1... Discriminator Loss: 0.3383... Generator Loss: 4.2185
Epoch 1/1... Discriminator Loss: 0.4962... Generator Loss: 1.6998
Epoch 1/1... Discriminator Loss: 1.1021... Generator Loss: 7.5592

Epoch 1/1... Discriminator Loss: 0.7536... Generator Loss: 0.9711
Epoch 1/1... Discriminator Loss: 0.1368... Generator Loss: 7.0807
Epoch 1/1... Discriminator Loss: 0.6029... Generator Loss: 0.8857
Epoch 1/1... Discriminator Loss: 0.2424... Generator Loss: 5.0247
Epoch 1/1... Discriminator Loss: 0.2617... Generator Loss: 1.7387
Epoch 1/1... Discriminator Loss: 0.1740... Generator Loss: 2.0405
Epoch 1/1... Discriminator Loss: 0.1183... Generator Loss: 5.3269
Epoch 1/1... Discriminator Loss: 0.1371... Generator Loss: 2.3507
Epoch 1/1... Discriminator Loss: 0.2864... Generator Loss: 2.0357
Epoch 1/1... Discriminator Loss: 0.1358... Generator Loss: 6.7378

Epoch 1/1... Discriminator Loss: 0.2119... Generator Loss: 2.9438
Epoch 1/1... Discriminator Loss: 0.1908... Generator Loss: 2.3017
Epoch 1/1... Discriminator Loss: 0.1538... Generator Loss: 4.9746
Epoch 1/1... Discriminator Loss: 1.6566... Generator Loss: 0.3177
Epoch 1/1... Discriminator Loss: 1.4480... Generator Loss: 0.4257
Epoch 1/1... Discriminator Loss: 1.5658... Generator Loss: 0.4749
Epoch 1/1... Discriminator Loss: 1.3108... Generator Loss: 2.8825
Epoch 1/1... Discriminator Loss: 1.3196... Generator Loss: 4.0045
Epoch 1/1... Discriminator Loss: 0.8684... Generator Loss: 0.7281
Epoch 1/1... Discriminator Loss: 0.7299... Generator Loss: 0.8661

Epoch 1/1... Discriminator Loss: 1.9081... Generator Loss: 0.5481
Epoch 1/1... Discriminator Loss: 0.7667... Generator Loss: 0.8577
Epoch 1/1... Discriminator Loss: 1.5340... Generator Loss: 0.4286
Epoch 1/1... Discriminator Loss: 1.2268... Generator Loss: 1.0560
Epoch 1/1... Discriminator Loss: 0.9150... Generator Loss: 1.0090
Epoch 1/1... Discriminator Loss: 1.2292... Generator Loss: 0.8261
Epoch 1/1... Discriminator Loss: 1.6194... Generator Loss: 0.3634
Epoch 1/1... Discriminator Loss: 0.9659... Generator Loss: 1.2710
Epoch 1/1... Discriminator Loss: 0.8187... Generator Loss: 2.1771
Epoch 1/1... Discriminator Loss: 1.0842... Generator Loss: 0.5904

Epoch 1/1... Discriminator Loss: 0.5876... Generator Loss: 2.4870
Epoch 1/1... Discriminator Loss: 0.6284... Generator Loss: 1.0814
Epoch 1/1... Discriminator Loss: 1.4151... Generator Loss: 2.2595
Epoch 1/1... Discriminator Loss: 1.5482... Generator Loss: 0.3783
Epoch 1/1... Discriminator Loss: 2.0301... Generator Loss: 0.2066
Epoch 1/1... Discriminator Loss: 1.4303... Generator Loss: 0.4340
Epoch 1/1... Discriminator Loss: 1.1183... Generator Loss: 0.5615
Epoch 1/1... Discriminator Loss: 1.0051... Generator Loss: 0.8940
Epoch 1/1... Discriminator Loss: 1.3086... Generator Loss: 0.4422
Epoch 1/1... Discriminator Loss: 1.2363... Generator Loss: 1.0030

Epoch 1/1... Discriminator Loss: 1.5778... Generator Loss: 2.3382
Epoch 1/1... Discriminator Loss: 0.8868... Generator Loss: 3.9566
Epoch 1/1... Discriminator Loss: 1.2611... Generator Loss: 2.0344
Epoch 1/1... Discriminator Loss: 0.9997... Generator Loss: 0.8256
Epoch 1/1... Discriminator Loss: 2.1116... Generator Loss: 0.2721
Epoch 1/1... Discriminator Loss: 0.9863... Generator Loss: 1.4487
Epoch 1/1... Discriminator Loss: 1.6329... Generator Loss: 0.4026
Epoch 1/1... Discriminator Loss: 0.6508... Generator Loss: 1.7059
Epoch 1/1... Discriminator Loss: 1.5193... Generator Loss: 1.5235
Epoch 1/1... Discriminator Loss: 1.3826... Generator Loss: 1.9684

Epoch 1/1... Discriminator Loss: 0.8962... Generator Loss: 1.4103
Epoch 1/1... Discriminator Loss: 1.2127... Generator Loss: 0.6719
Epoch 1/1... Discriminator Loss: 1.3663... Generator Loss: 0.4246
Epoch 1/1... Discriminator Loss: 1.1544... Generator Loss: 0.6655
Epoch 1/1... Discriminator Loss: 1.7973... Generator Loss: 0.2282
Epoch 1/1... Discriminator Loss: 1.3549... Generator Loss: 0.4011
Epoch 1/1... Discriminator Loss: 0.9909... Generator Loss: 1.1155
Epoch 1/1... Discriminator Loss: 0.8815... Generator Loss: 2.2619
Epoch 1/1... Discriminator Loss: 1.9800... Generator Loss: 0.2730
Epoch 1/1... Discriminator Loss: 2.1945... Generator Loss: 0.1468

Epoch 1/1... Discriminator Loss: 1.5421... Generator Loss: 0.3388
Epoch 1/1... Discriminator Loss: 1.3466... Generator Loss: 0.7026
Epoch 1/1... Discriminator Loss: 1.4293... Generator Loss: 1.0913
Epoch 1/1... Discriminator Loss: 1.4787... Generator Loss: 0.5693
Epoch 1/1... Discriminator Loss: 1.3243... Generator Loss: 1.0842
Epoch 1/1... Discriminator Loss: 1.1571... Generator Loss: 0.7459
Epoch 1/1... Discriminator Loss: 1.4020... Generator Loss: 0.4724
Epoch 1/1... Discriminator Loss: 1.1820... Generator Loss: 0.9604
Epoch 1/1... Discriminator Loss: 1.0169... Generator Loss: 1.4610
Epoch 1/1... Discriminator Loss: 1.7429... Generator Loss: 0.2555

Epoch 1/1... Discriminator Loss: 1.4992... Generator Loss: 0.5118
Epoch 1/1... Discriminator Loss: 1.0774... Generator Loss: 1.2635
Epoch 1/1... Discriminator Loss: 1.4813... Generator Loss: 0.4725
Epoch 1/1... Discriminator Loss: 1.9656... Generator Loss: 0.2206
Epoch 1/1... Discriminator Loss: 1.0940... Generator Loss: 0.6655
Epoch 1/1... Discriminator Loss: 1.7473... Generator Loss: 0.3405
Epoch 1/1... Discriminator Loss: 1.0815... Generator Loss: 0.9496
Epoch 1/1... Discriminator Loss: 1.7022... Generator Loss: 0.3374
Epoch 1/1... Discriminator Loss: 1.1639... Generator Loss: 1.5708
Epoch 1/1... Discriminator Loss: 1.8346... Generator Loss: 0.2296

Epoch 1/1... Discriminator Loss: 1.2985... Generator Loss: 0.5059
Epoch 1/1... Discriminator Loss: 0.9840... Generator Loss: 0.7880
Epoch 1/1... Discriminator Loss: 2.4241... Generator Loss: 0.1171
Epoch 1/1... Discriminator Loss: 1.6958... Generator Loss: 0.2898
Epoch 1/1... Discriminator Loss: 1.2291... Generator Loss: 1.7416
Epoch 1/1... Discriminator Loss: 1.2677... Generator Loss: 0.4580
Epoch 1/1... Discriminator Loss: 1.9798... Generator Loss: 0.2307
Epoch 1/1... Discriminator Loss: 1.0211... Generator Loss: 0.9819
Epoch 1/1... Discriminator Loss: 1.4500... Generator Loss: 0.3582
Epoch 1/1... Discriminator Loss: 0.8845... Generator Loss: 0.8427

提交项目¶

提交本项目前，确保运行所有 cells 后保存该文件。

保存该文件为 "dlnd_face_generation.ipynb"，并另存为 HTML 格式 "File" -> "Download as"。提交项目时请附带 "helper.py" 和 "problem_unittests.py" 文件。